Tools for Practice

#130: Working out depression: Is exercise effective for depression?

Can exercise improve mild to moderate depression?

Exercise shows benefit in treating mild to moderate depression but the evidence is at a high risk of bias with smaller benefit in higher quality studies. In high quality studies showing benefit, approximately one in 8-12 patients with mild-moderate depression will attain remission due to exercise.   

CFPCLearn Logo

Reading Tools for Practice Article can earn you MainPro+ Credits

Join Now

Already a CFPCLearn Member? Log in

At least 10 systematic reviews have been done.1-10 Focusing on four of highest quality in depression [13-58 randomized controlled trials (RCTs), 720-2,982 participants, usually mild-moderate depression]:1-4 
  • Results in standard mean differences (SMD): A statistical tool to combine different scales with limited clinical meaning.   
  • Effect on depression compared to no treatment or control: Three meta-analyses found statistically significant improvements in depressive symptoms, ranging from “moderate” to “large” effects (SMD 0.6 to 1.1).1,2,4 Results inconsistent.  
    • In high quality studies:  
      • Using randomization concealment, intention-to-treat and blinded assessors: Two meta-analyses were no longer statistically significant.1,2  
      • Using only publication in peer reviewed journal/dissertation: Remaining analysisSMD reduced from 1.1 to 0.7 (now a “moderate” effect).4 
    • Duration reduced effect from “large” (SMD=1.8) in studies <8 weeks duration, to “moderate” (SMD=0.6) if >8 weeks.4 
    • Exercise not statistically different from psychotherapy (seven RCTs, 189 patients) or medication (four RCTs, 300 patients).1 
    • Limitations: Many studies used patient self-report for results;1,4 small sample sizes; exercise was usually supervised/done in groups (therefore socializing could effect results);4 blinding, randomization concealment, and intention-to-treat infrequent. 
  • Examining the five highest quality RCTs:11-15 Three of five found statistically significant (or nearly): 11,12,15 
    • Response (≥30% reduction in Hamilton Depression score): Number Needed to Treat (NNT) 5 over 10 weeks.15 
    • Remission (“normal” Hamilton Depression score): NNT 8-12 over four months.11 
  • Exercise has dose-dependent, positive effects on quality of life in non-depressed patients.16  
  • In chronic illness without depression, exercise has an inconsistent, small-moderate effect on depressive symptoms (SMD=0.3).9 
  • Largest effects of exercise were observed among those with mild-to-moderate depression.8 
    • Evidence is insufficient to judge if one type of exercise is better.3,4 

This content is certified for MainPro+ Credits, log in to access


  • Adrienne J Lindblad BSP ACPR PharmD
  • Doug Klein MD CCFP MSc
  • Shweta Dhawan BSc

1. Cooney GM, Dwan K, Greig CA, et al. Cochrane Database Syst Rev. 2013; 9:CD:004366.

2. Josefsson T, Lindwall M, Archer T. J Med Sci Sports. 2014; 24(2):259-72.

3. Rethorst CD, Wipfli BM, Landers DM. Sports Med. 2009; 39(6):491-511.

4. Lawlor DA, Hopker SW. BMJ. 2001 Mar 31; 322(7289):763-7.

5. Mura G, Carta MG. Clin Pract Epidemiol Ment Health. 2013; 9:125-35.

6. Bridle C, Spanjers K, Patel S. Br J Psych. 2012; 201(3):180-5.

7. Blake H, Mo P, Malik S, et al. Clin Rehabil. 2009; 23(10):873-87.

8. Sjosten N, Kivela SL. Int J Geriatr Psychiatry. 2006; 21(5):410-18.

9. Herring MP, Puetz TW, O'Connor PJ, et al. Arch Intern Med. 2012; 172(2):101-11.

10. Mura G, Moro MF, Patten SB, et al. CNS Spectr. 2014 Dec; 19(6):496-508.

11. Blumenthal JA, Babyak MA, Doraiswamy PM, et al. Psychosom Med. 2007 Sep-Oct; 69(7):587-96.

12. Blumenthal JA, Sherwood A, Babyak MA, et al. J Am Coll Cardiol. 2012 Sep 18; 60(12):1053-63.

13. Hoffman JM, Bell KR, Powell JM, et al. PM R. 2010 Oct; 2(10):911-9.

14. Krogh J, Saltin B, Gluud C, et al. J Clin Psychiatry. 2009 Jun; 70(6):790-800.

15. Mather AS, Rodriguez C, Guthrie MF, et al. Br J Psychiatry. 2002 May; 180:411-5.

16. Martin CK, Church TS, Thompson AM, et al. Arch Intern Med. 2009; 169(3):269-278.

Authors do not have any conflicts to disclose.